Stressing stress
Angela Kennedy said: Does the 0.21 refer to relative risk? And if so, are they talking 0.21 relative risk overall of ANY 'psychological' variable included together as a category? The reason I say that is because they seem to be saying in the abstract that there were correlations with only some variables?
I haven't accessed the full text either so I don't know exactly what statistical method they used. I'm not good at statistics anyway but I'm guessing their "effect size correlation coefficient" would have a range of 0 to 1, with 0 being no correlation and 1 being perfect correlation, while 0.2 = small, 0.5 = medium, 0.8 = large. AFAIK an effect size of 0.21 is considered small no matter what method they used.
Falagas et al 2010 say "significant overall main effect of psychological stress" and I don't know how this was measured or if they lumped all types together, but the abstract does say "effect sizes for the association did not vary according to type of stress". We don't really know what other factors are involved.
biophile said: My critique style of "stress and disease" so far has mostly been limited to arguing that relatively weak associations don't justify the hyped psychobabble and have also been reported in dozens of diseases that are not usually thought of as "stress-related illnesses". It will take a long time to look into deeper into this sort of research to determine whether even this approach has been unsafe due to methodological artifacts, the whole field may be in question.
Esther12's vs Angela Kennedy's interpretation of what I said]
oceanblue said: there is clearly still some doubt as to whether or not the associations are correct
Angela Kennedy said: [...] any 'evidence' used to support claims of 'stress' causing or perpetuating illnesses is highly unsafe, because of various methodological errors in the research, and theoretical errors around how to define 'stress'. The case for 'stress' causing or perpetuating all sorts of illness or health problems has been grossly overstated at best.
Esther12 said: I agree that measuring 'stress' etc is difficult, and this this sort of research could easily be distorted by the presumptions of those involved or the various other biases which can distort medical papers... but I don't think that means we should assume all the research linking stress/etc to health problems is likely to be wrong. Our brains are mighty resource hogs, and it's difficult to assume that prolonged and extreme psychological strain would have no impact upon the rest of our bodies, especially given the role our brains play in regulating so many of our our bodily functions.
You both were correct in your interpretations. Let me elaborate on my position.
It is often claimed that CFS involves increased comorbidity with other functional syndromes, high rates of and association with psychiatric comorbidity, "stress-related", psychopathology can influence reporting of symptoms, etc. These are often implied as or alluded to by the psychobabble as somehow unique to psychosomatic and psychiatric diagnoses, or evidence for psycho>somatic mechanisms in the supposed absence of recognised pathophysiology.
However, there is a mountain of reported evidence for such phenomenon in many medical diseases as well (often the evidence is conflicted and/or there are systematic reviews which sober up all the hyperbole), so all these claims are probably a form of special pleading when applied to CFS. One caveat here is that when compared to other medical diseases there does seem to be a trend for somewhat stronger associations for some aspects of so-called medically unexplained and functional symptoms in some scenarios but again there are major methodological issues here as well that need resolution.
The correlations found in research on psychological factors in illness/disease often do survive after adjusting for obvious confounding factors, sometimes they do not or are reduced. But the effect sizes, relative risks, odds ratios, hazard ratios etc in these studies aren't all that impressive either, usually nothing like smoking and lung cancer, and rarely of clinical value (unless severe or the disease already exists).
Issues with defining and measuring "stress" or other psychological and social variables is another problem, Angela has looked into this much more than me, I have only begun to investigate that (slowly like a turtle!). Until then, I will continue to tentatively accept there are probably some genuine associations between comorbid and premorbid psychological dimensions and medical disease, but I believe the exaggeration and overgeneralisation we often see in the common psychobabble simply isn't justified by the evidence, while the reported links from research can't be routinely taken at face value either. This of course doesn't mean that stress isn't relevant to one's health.
I also consider possible alternative explanations for such associations, like shared risk factors, overlapping pathophysiology, early psychological manifestations/vulnerabilities of disease pathology that are misinterpreted as causal when physical symptoms follow later, disturbed homeostasis and neuroendocrinimmunological overlap with disease pathology giving the illusion of a "mind over body" mechanism (which wouldn't exist without the primary disease), etc.
Then, as Angela said, "not to mention psychologist's fallacies in deeming negative appraisals of predicaments 'maladaptive' or 'irrational'". I think even the role of psychosocial factors in the onset of psychiatric diagnoses may be overstated as well. Also, I noticed that some people view "chronic whiplash" and "chronic lower back pain" as obvious psychosocial problems, but again, systematic reviews into psychosocial factors don't really support this psychosocialisation either. This isn't to say such factors don't exist, but to note the theme of blanket hyperbole.
I heard an anecdote about someone who goaded suspected "somatisers" into getting angry as a demonstration in order to help them realise that their worsening symptoms were evidence their "illness" was primarily rooted in emotions. Some illnesses/diseases may not respond abnormally to anger but that certainly isn't universal and is a rather stupid litmus test. What do they think would happen if they angered someone who just a heart attack, or a stroke, or are suffering the consequences of chemotherapy, or are in emergency care, or have a significant infection, or are experiencing migraine or sleep deprivation, or have one of many other conditions which leaves the patient as impaired as what is often experienced in ME/CFS?
Yep, particularly the parts about overstatement, difficulties unraveling the numerous forms of different kinds of stress (negative, positive, "physical", "psychosocial"), conflating mind with brain, black boxing and lack of demonstrated mechanisms, lack of critique towards dubious claims regarding stress in somatic illness, inappropriately blaming patients for any associations, us not affording the luxury of reams of faulty psychobabble just in case a small portion may one day be correct. There is a systematic review (
Guiraud et al 2010) which finds both "negative or positive emotions" are significant risk factors for triggering ischemic stroke, do we all now need CBT to correct faulty positive thinking?
WillowJ said: another thing that was pointed out by biophile is that the amount of risk attributed to stress seems to be small. in a complex system such as real life, we can't control all the variables. It isn't always the variable we thought we were testing in the experiment, that produced the effect. especially when the result is small, we should be careful about drawing conclusions. but that's not what has been done with the stress idea. stress is a well-accepted idea which is thought to be a major contributor (or even the main factor) in a number of diseases, but we really don't have compelling evidence for that.
Indeed.